Ben Radack 🏝️

11K posts

Ben Radack 🏝️

@benradack

Media buyer & creative strategist - I help brands lower their CPA with Facebook (Meta) ads

Los Angeles, CA Katılım Eylül 2020

586 Takip Edilen17.2K Takipçiler

Sabitlenmiş Tweet

Ben Radack 🏝️@benradack·28 Ara

I took over this account 2 days ago and dropped their CPA by 74%. All with a SUPER simple account structure & testing strategy Creative Testing - ABO → Went through their ad library (hundreds of ads) → Batch tested a ton of them across 5 different ad sets → Yes, Meta skipped a bunch, but it helped me find winners fast Scaling Campaign - Cost Cap CBO → Ad set 1: All historical winners from the account → Ad set 2: All proven winners from December testing - min spend limit That's it. What was broken: Their past team was hardly testing anything. All the spend was going to 1 ad that was extremely fatigued. The account had zero velocity. The fix: Simplify the account structure + increase testing velocity significantly. Sometimes the best strategy is just testing more and organizing better. You don't need 10 campaigns and complex structures. You need clean separation between testing and scaling, and you need to actually test. 2 days in, CPA dropped 74%.

English

168

24.3K

Ben Radack 🏝️@benradack·9h

@MenachemAni What. How.

English

142

Menachem (Google Ads)@MenachemAni·9h

Got a lead from Claude today

English

427

Ben Radack 🏝️@benradack·1d

If you're scaling and testing in the same CBO, make sure you set max spend limits on your testing ad set. Here's why. Once in a while you'll find a new test that takes all your campaign spend and still fails. By setting a max spend limit at around 20% of your overall budget, you keep that ad from running away. Min spend limits force Meta to deliver on tests. Max spend limits stop unproven ads from blowing up your day. Why not just test in a separate ABO? You can. But like I said in my last tweet, I both trust the algo and don't. By keeping testing inside a CBO with a wide gap between min and max spend limits, I'm letting Meta still pick where to put spend. If it keeps pushing my testing ad set and the tests are crushing, I just raise the max limit. That's me trusting the algo when it's earned it. If a test isn't proving out, the max limit caps the damage. That's me not trusting the algo when it's making a bad call. Min and max spend limits give you a sliding scale of control over Meta's delivery. Use them to manage how much you let the algo run the show.

English

1.4K

Ben Radack 🏝️@benradack·9h

@herrmanndigital You sure it’s not Peter Thiel? Lol

English

283

David Herrmann@herrmanndigital·10h

I think Sam Altman is the Antichrist

Vivek Sen@Vivek4real_

SAM ALTMAN: “WE SEE A FUTURE WHERE INTELLIGENCE IS A UTILITY, LIKE ELECTRICITY OR WATER, AND PEOPLE BUY IT FROM US ON A METER.”

English

13.3K

Ben Radack 🏝️@benradack·12h

@oliverwhudson Format is targeting. same angle, different vehicle, different pocket of the audience. five shots instead of one

English

183

Olly Hudson@oliverwhudson·16h

We scaled one angle across 5 vehicles and drove £90K+ with a framework database Hit rate moved from 4% - 10%. Peaked at 18% on that product Here's what the before and after looked like: Before: winning angles tested in one format, noted, abandoned. No horizontal scaling across vehicles or personas After: one proven angle scaled across 5 vehicles. Format = targeting. 5x the chances to find the soulmate ad. --- Before: Inspiration pulled on instinct. Big swings based on gut feel. High risk, low conversion on net-new concepts. After: shifted from "inspo I like" to "inspo with real engagement." De-risked every concept before it entered production. --- Before: no systematic breakdown of why top ads worked. Surface level angle analysis without touching the psychology underneath. After: 90 day component breakdown. Every winner deconstructed by valence zone, awareness level, vehicle, and persona. Pattern recognition identified as the core psychological driver. --- The framework database is what connected the dots. Without it, a winning angle gets noted and abandoned, but with it one proven message becomes a production system

Olly Hudson@oliverwhudson

One of the most underutilised assets in any DTC brand is the framework database Every ad that works is a hypothesis and a claim about why this angle, this persona, this emotional register converts. Most brands dont log that, they only note the metric, ship the next batch, and lose the learning We automate it: any ad that hits £1,000 in spend gets pulled automatically, uploaded to Gemini, analysed for framework, visual structure, shot type, and hypothesised reason for performance. It goes into the database When the script writing skill runs, it has access to every one of those frameworks and it picks the one most aligned to current data, gives a hypothesis for why, and builds the brief from there If you don't have that infrastructure yet, start manually. Go to Meta Ads Library, sort by highest impressions for your competitors, download, upload to Gemini, ask for framework breakdown and hypothesis, build the database one asset at a time That database becomes your unfair advantage. The AI that runs on top of it gets better every time a new winner is logged

English

Ben Radack 🏝️@benradack·13h

@rokhladnik Turn off the retargeting campaign for a week. if blended cpa stays the same, you know those conversions were happening anyway.

English

217

Rok Hladnik@rokhladnik·15h

A 0.5x ROAS prospecting campaign might be doing exactly what it should A 6x ROAS remarketing campaign might be doing almost nothing That’s why blended numbers matter Funnel stage matters New customer share matters Incrementality matters ROAS without context is just a number The same ROAS can mean two completely different things depending on where it came from

English

870

Ben Radack 🏝️@benradack·13h

@IstvanicMarin the discount ad running at 2.4x frequency tells everything. separate retargeting campaign would've just bid against the same people and made the numbers look better than they are

English

243

Marin.Istvanic@IstvanicMarin·16h

You DON’T Need TOF/MOF/BOF Campaigns Here’s what’s actually happening with Facebook’s targeting + retargeting logic: Let’s break down two ads that are crushing it inside my ASC scaling campaign. The top spender is a short mini‑VSL. Its frequency sits at 1.04, so it’s mostly being served to fresh eyes & and the audience breakdown confirms it’s reaching new customers. Meanwhile, there’s a simple image ad pushing a 40% off offer. It spent about 2.5x less, but it’s running at roughly double the frequency, meaning it’s largely landing on warm traffic and people who’ve already interacted. And that’s the punchline: Facebook is already assigning the “right” audience to each ad automatically, without you carving up TOF/MOF/BOF buckets by hand. Which is exactly why a separate retargeting campaign is redundant.

English

6.9K

Ben Radack 🏝️@benradack·15h

@herrmanndigital I've launched a couple new ad accounts for my own projects recently and the first thing it does is spend most of the budget on the audience network. standard is to have it shut off now

English

535

David Herrmann@herrmanndigital·16h

Since removing audience network on Meta completely for a brand our CVR is up 142.52% on landing page views. Crazy enough it's also led to our overall frequency decreasesing, CPMr decreasing. My proxy for when determining when to pause audience network entirely is as follows: 1. Is your CTR somewhat high? You'll know because you'll see audience network getting 10-15-20-30-40% CTRs. 2. Is your audience network getting at least 1% of daily budget? 3. Has your CVR on landing page views to purchases stayed flat or relatively flat for the year?

English

184

24.9K

Ben Radack 🏝️@benradack·18h

Today I'm taking over a new account spending $200k/month. Here's what I'm doing on day 1. **First: Launch all 40 new concepts (80 ads)** 4 ads per ad set, one ad set per concept. We can't afford to test each concept in a separate ABO and wait weeks for data. So it's all going straight into the scaling campaign. **Next: Dig through all previously tested ads** Run those in the background through an ABO. Force spend into the concepts that never got delivery the first time. Probably a lot of missed winners sitting in there. **Then: Launch a historical winners campaign** Old ads that haven't run in months. Bringing them back is one of the fastest ways to turn performance on. The goal on day 1 is simple. Find winners. Stop the bleeding. Once we do that and get things under control, I'll go back to a more methodical testing approach. But right now speed matters more than precision. You can clean up the account later. First you have to stabilize it.

English

1.4K

Ben Radack 🏝️@benradack·1d

@IstvanicMarin abo for testing, cbo for scaling proven ads. the accounts that struggle are the ones trying to do both in the same place

English

301

Marin.Istvanic@IstvanicMarin·1d

Two ecom founders recently told me they were honestly shocked my ABO campaigns are hitting like this. Makes sense. Sounds like their last agency went all-in on the “just let Meta cook” CBO philosophy. So Facebook starts funneling spend into whatever ad sets win on: - thumb-stops or - CTR even if the CPA is quietly bleeding out in the background. And yeah, I know, ABO is more hands-on. I’m fine paying that price if it buys better performance. Big agencies aren’t. That’s the whole gap.

English

2.5K

Ben Radack 🏝️@benradack·1d

@heyitsalexP at a certain scale the ad stops being just about conversion. it's also the first impression for millions of people who will never click. that math changes completely

English

298

Alex@heyitsalexP·1d

"doesn't matter if the ad is ugly as long as it scales" this post has almost 700k impressions. if you're a Chewy-scale brand, you can't run the same ad creative as a drop shipping brand...

kp⁷💜 is seeing bangtan@bulletproofkp

Just curious @Chewy twice now I’ve gotten ads on my instagram feed from your blue check verified instagram page and it’s of a quarter with maggots on it I’m so confused Is AI running your instagram ad campaigns and simultaneously ruining them?

English

6.6K

Ben Radack 🏝️@benradack·1d

@KodyNordquist It seems like a media buying issue. They are aware of hte image but looks like it could be a DPA pulling in imagery from somewhere.

English

174

Kody Nordquist@KodyNordquist·1d

Chewy is running a 101 on how not to do creative.

kp⁷💜 is seeing bangtan@bulletproofkp

English

1.1K

Ben Radack 🏝️@benradack·1d

@bulletproofkp @Chewy Bad media buying

English

391

kp⁷💜 is seeing bangtan@bulletproofkp·2d

English

6.3K

Ben Radack 🏝️@benradack·1d

Your trust in Meta's algorithm dictates how you manage your account. If you're like me, you neither fully trust nor fully distrust the algo. So I use CBO campaigns to scale because Meta is good at distributing budget across proven ad sets/ads. But I use ABO campaigns to test because I don't trust Meta to not skip over winners during testing. If you fully trust the algo you probably scale and test in the same campaign. Maybe even the same ad set. Let Meta figure it out. I don't operate that way because I've seen what can happen. → Meta picks the wrong ads to spend on → Meta keeps spending on fatigued ads way longer than it should → I've let the algo do all the work and watched it dump my entire campaign budget into a new ad that ended up failing Thousands wasted before I caught it. That said, no one approach is "right." What works for you is your truth. Some accounts perform better when you control everything. Some perform better when you let Meta drive. The trick is knowing where you stand on this and building your account around it.

English

2.1K

Ben Radack 🏝️@benradack·2d

Creative diversity is literally this and more. Not making the same ad every time Because when you make different looking ads with different messaging or even simple changes like photos or imagery you hit different pocket audiences And since Meta does the targeting on the ad level, that becomes more important than ever Here's how I look at it Some people are more likely to purchase from ads that have gold lettering And some aren't So the ad with the gold lettering hits that small little pocket audience. A more obvious example would be simply calling out a group of people. Since Meta reads your ads, it knows who you're trying to target.

Nick Theriot@nicktheriot_

Creative diversity checklist I use to scale brands to $1M+/month: 1. Photos AND videos (not just one) 2. Scripted UGC 3. Non-scripted UGC 4. IG Reel trends 5. Professional shoots 6. Long-form VSLs 7. Testimonial photos 8. Product + benefit photos 9. Before/after photos 10. Press/news-style graphics 11. Native ads Bookmark this.

English

Ben Radack 🏝️@benradack·2d

@nicktheriot_ This is a good post.

English

641

Nick Theriot@nicktheriot_·2d

English

180

16.2K

Ben Radack 🏝️@benradack·2d

Every ad you run hits a different pocket audience. Some ads have small pockets. They work great at $200/day and fall apart at $600. That's because the audience that converts off that ad isn't big enough to absorb the higher spend. Some ads have huge pockets. They can scale to $5,000+/day without CPA shifting much because the audience that responds to their angle is massive. These bigger pocket ads are the ones you're really trying to make. They're the ads that move the needle and let you scale the account. This is also what ad fatigue actually is. It's not about frequency. It's not about your ad getting "old." It's about running out of audience for that specific angle. When the people who would have responded to that hook have all seen it, you're done. Two things follow from this: → Not every winning ad has scale. Some are pocket winners. Let them run at their natural budget instead of forcing them to do more. → The ads that scale the highest are the ones speaking to the broadest part of your market. Test wider angles, not just variations of what already works. You don't fight ad fatigue by adjusting settings. You fight it by finding ads that can access bigger audiences.

English

1.8K

Ben Radack 🏝️@benradack·2d

One of the most important metrics to track is your cost per add to cart. ROAS is a lagging indicator. By the time it tells you an ad is working, you've already wasted days of spend deciding. Cost per ATC tells you in hours. Why it matters. → Add to carts come way before purchases. Especially on high AOV products where buyers take days to convert. → A low cost per ATC means the ad is bringing people into the funnel cheaply. That's the hardest part of the job. → A high cost per ATC means the ad isn't connecting. No amount of waiting fixes that. What I do. I watch cost per ATC on every new ad I launch. If it's strong by day 2-3, I give the ad more runway even if purchases haven't come in yet. If cost per ATC is bad, I know the ad isn't going to suddenly fix itself when purchases catch up. Stop waiting for ROAS to tell you what cost per ATC already knows.

English

1.9K

Ben Radack 🏝️@benradack·3d

@martinlofvall Sure! That can work. Do you mean how you judge your ads?

English

Martin@martinlofvall·3d

@benradack Cost cap at 7 day avg. cpa?

English

102

Ben Radack 🏝️@benradack·3d

You can run your entire Meta ad account with one CBO. → Ad set 1: Historical winners with a cost cap → Ad set 2: Creative tests with min and max spend limits How the spend limits work. Min spend limit on ad set 2 forces your new tests to get budget alongside your old winners. Without it Meta keeps dumping spend into the historical ad set and ignores everything new. Max spend limit on ad set 2 keeps a new test from running away if Meta suddenly favors it before you've validated it. Unproven ads can spike CPA fast if you let them eat the whole campaign. This setup works for accounts spending under $1k/day. Simple structure, forced delivery on new ads, easy to manage.

English

2.9K

Ben Radack 🏝️@benradack·3d

@UgcSam It’s based mainly on the campaign budget. If it’s set at $3k. Then I don’t want the new ads to take up all the spend so I’ll probably put the max at 30%

English

Sammy D@UgcSam·3d

@benradack how do you deicde what to set the max at?

English

Keşfet

@MenachemAni @herrmanndigital @oliverwhudson @rokhladnik @IstvanicMarin @heyitsalexP @KodyNordquist @elonmusk